Philosophy Majors and the GRE: Updated Data (w/updates)


When students are compared by major on how far above average they do on the Graduate Record Examinations (GRE), a standardized test used in many disciplines to assess applicants to graduate programs, philosophy majors come out on top, according to a new look at test score data over the past few years.

Tomas Bogardus, associate professor of philosophy at Pepperdine University, had noticed that much of the data easily available about how philosophy undergraduates fare on the GRE (such as on the Value of Philosophy pages) was from 2015 at the latest, and helpfully compiled newer information, including the chart above and the table below. [NOTE: See the updated versions of the above chart, with more majors, in the updates at the end of this post.]

He writes:

If you just add up raw scores from the three sections, Philosophy doesn’t have the highest total score. But, because of the different standard deviations for each section, you might think that what’s more interesting is how far above average students are scoring on each section, i.e. how many standard deviations above the mean for all students is the mean score for each major. When you do that, and take the average number of standard deviation from the mean for all three sections, Philosophy majors are on top. And that seems to be because of how exceptionally well they do on the Verbal and Writing sections, which makes up for their relatively modest (but still well above average!) score on the Quantitative section. 

Below is a table showing mean GRE scores, by major. In addition to performing better than all others in terms of standard deviations above the mean, philosophy majors score higher than all others on the Verbal and Writing sections of the exam. They also score higher than all other humanities majors on the Quantitative section of the exam. (Note: the following table is a revised version of the one originally included in this post. Thanks to Tom Hurka for spotting a problem with the original table.)

I’ve updated the Charts & Graphs page with this information. They also appear on Professor Bogardus’ own “Why Study Philosophy?” page, which readers should check out.

UPDATE: Here’s the data on which the above is based (note: the following spreadsheet is a revised version of the one originally included in this post):

UPDATE 2: Professor Bogardus made a new chart that includes more of the majors:

And here is the same information presented differently:


guest
34 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
not clever
not clever
1 year ago

I feel like an idiot for asking this but here goes:

How is it that, on average, pretty much every major scores above the mean? What majors are scoring below the mean? Report

Dale E Miller
Reply to  not clever
1 year ago

The testing center is in Lake Woebegone. Report

Alastair Norcross
Reply to  not clever
1 year ago

A couple of guesses: (i) some of the majors here that do score below the mean are very large majors; (ii) there are quite a few majors not listed here.Report

Tom Hurka
Tom Hurka
1 year ago

I don’t understand the Writing (Converted) scores, where higher gross scores seem to result in lower converted scores.

Report

Dale E Miller
Reply to  Tom Hurka
1 year ago

Can you give an example of a couple majors for which this is true? I’m not seeing it.Report

Dale E Miller
Reply to  Dale E Miller
1 year ago

Oh, because it’s been fixed.Report

Tomas Bogardus
1 year ago

Not Clever: That’s a good question. Professor Norcross is right: I didn’t include all the majors in the data table. Really, just the ones I thought were most popular. Also, on the chart, I didn’t include majors with negative values for distance from the mean, because it made the chart pretty cluttered and messy. (Somewhat regrettable, since it might have been useful to have Business Admin appear on the chart…) If you look at the data table, you’ll see some majors that are below the mean. And if you check the complete data sets, you’ll see even more majors.

Tom Hurka: You’re right! Thanks for catching that! I had input the formula incorrectly to convert the Writing score to a 130-170 point scale. (My bad! I had it right on a second sheet, which I didn’t include with the Excel spreadsheet I’d sent Justin. But then I just flat out input it wrong. Mea maxima culpa.) I’ve sent Justin a corrected version.

You can see the image of it here: https://www.dropbox.com/s/kqkb5f2ruhmaxho/GRE%202019%20table.png?dl=0

And you can access the Excel Spreadsheet here:
https://www.dropbox.com/s/rb3fkknb50lu7d5/GRE%20Data%20%28no%20graph%29.xlsx?dl=0

So, even with the correction, Philosophy still is not on top with the combined raw score (looks like we’re #2 behind Physics&Astronomy), but this error didn’t affect the average #SD from the mean calculation at all. Philosophy is still #1 there.

Thanks again!Report

Tomas Bogardus
1 year ago

It seems like there might be some interest in a chart that includes more majors, including ones like Business Administration, as well as Accounting. The result is a little less pretty, and the font is definitely smaller, but in case you might find it useful, I’ll include it here:

https://www.dropbox.com/s/ecqbw8wii810d0w/GRE%202019%20second%20chart.png?dl=0

Report

Brian Weatherson
Brian Weatherson
Reply to  Tomas Bogardus
1 year ago

Can you do it with the majors on the y-axis and the scores on the x-axis, so the fonts don’t have to be so tiny?Report

Tomas Bogardus
Reply to  Brian Weatherson
1 year ago

Good idea! This is stretching my MS Excel skills to the limit, but how does this look? 🙂

https://www.dropbox.com/s/76x5ghr2nj2u8c5/GRE%202019%20third%20chart.png?dl=0
Report

Outlander
Outlander
1 year ago

It’s probably worth clarifying that the major grouping is based on intended graduate major and not undergrad major (though there’s overlap obviously) – see description at top here https://www.ets.org/s/gre/pdf/gre_guide_table4.pdf. That is, a philosophy major who intended to go to law school would be filed under the “law” bin in this data. Report

Tomas Bogardus
Reply to  Outlander
1 year ago

Oof, yes, good point! I hadn’t thought of that. Definitely worth clarifying. Though one might think there won’t be tooooo many law-school-bound philosophy majors taking the GRE, no? I’d think they’d probably be taking the LSAT… But, yes, definitely worth pointing out! I’m not aware that ETS keeps track of the actual majors of their test takers, which would be nice data to have. Report

grad student
grad student
Reply to  Tomas Bogardus
1 year ago

Curious–if that’s the case, then what’s wrong with this simple explanation. Philosophy graduate admissions is highly competitive and a bit niche, and so the vast majority of philosophy majors applying to graduate philosophy programs are very strong students. Very strong students tend to better on the GRE. Compared to, for example physics and astronomy, I would guess (though I could be wrong) that you don’t need to be a top student in order to gainfully apply to graduate school.

This also explains why more humanities (religion/theology, english, foreign languages) appear at the top. I would guess (though again I could be wrong) that a majority of those applicants are either applying to PhD programs or to master’s with an interest in a PhD, and hence those students will be fairly strong. Terminal MAs in philosophy are a thing, but from what I understand most are geared towards the PhD. But there are more, e.g. Biological/Biomedical sciences graduate programs where a PhD or other highly competitive program is not the aim. In other words, there are useful programs for weaker students in the sciences, but there are very few for the humanities.

Are there any interpretations of the data that suggest that this isn’t what’s going on? I have no agenda, just curious. Report

Pendaran
Pendaran
Reply to  grad student
1 year ago

Yes, this is what I was thinking. Another factor is the relative importance that the fields place on the GRE. Some fields rely on the test more so than others. The more competitive a field is and the more they weight the GRE the higher the GRE scores we should expect from the majors. Another related factor is the type of person who is applying to graduate school in that discipline and whether they are going for a PhD or a MA. It’s irresponsible to use these data to suggest that a BA in philosophy causes you to perform better on the GRE. Report

Pendaran
Pendaran
Reply to  Pendaran
1 year ago

I should have said this:

The more competitive a field is and the more they weight the GRE the higher the GRE scores we should expect from the applicants to that field. Report

Tom Hurka
Tom Hurka
1 year ago

It’s ages ago now, but back in 1989 I wrote a newspaper column about the then-version of these statistics, which the APA was publicizing, which also included the MCAT, GMAT, and LSAT. As I recall, the breakdown then was by undergrad major, not intended later subject of study, and the philosophers had the highest score on the LSAT as well as on the GRE verbal. The APA take-home was that philosophy students had the highest lowest rank of students from any discipline, i.e. their rank on the test they did worst on (GRE quantitative) was higher than the rank on their worst test by any other group, e.g. the physics students ranked lower on the GRE verbal than the philosophy students did on the GRE quantitative.

The headline my editor put on the column was “How to Get to the Top: Study Philosophy”, which I thought was nice. As others have noted, philosophy students’ performance on these tests has been consistent for many decades now.Report

Tomas Bogardus
1 year ago

Here’s a breakdown by subsection, in case it might be useful:
https://www.dropbox.com/s/w5djo50soc5f701/GRE%20subsections%20chart.png?dl=0Report

Wayne Fenske
Wayne Fenske
1 year ago

Presumably, the point of compiling this data is to convince people that if they choose to major in Philosophy rather than, say, Psychology, they will end up with better cognitive skills – as evidenced by the difference in GRE scores for Philosophy majors as compared to Psychology majors.
That may be so, but these GRE test score data don’t show that it is because they only give us subject variable comparisons.
The experimental protocol required to establish a causal connection between choosing to major in Philosophy and having better cognitive skills, as reflected in higher GRE scores, would call for taking a large number of people, randomly assigning them to complete various undergrad majors, then getting all of them to take the GRE, and only then comparing the correlations between someone’s GRE score and their undergrad major.
If the Philosophy majors stood out under these experimental conditions, then the claim that one can improve their cognitive skills by doing a Philosophy major instead of a Psychology major would indeed be established.
Alas, from a practical point of view, such a properly orchestrated experiment cannot be carried out, of course.
As it stands, it could be that the positive correlation between majoring in Philosophy and possessing the superior cognitive skills that enable one to achieve an above average GRE score is best explained by the superior training one’s receives in the course of completing the requirements for a Philosophy degree.
But it could also be that the type of person who is attracted to the prospect of majoring in Philosophy will tend to have above average propensities for abstract and linear thought, mental compartmentalization, structured understanding, etc. as well as healthy supplies of mental curiosity, bookishness, introversion, etc.
If this is so, then the best explanation of a person both choosing to major in Philosophy and scoring above average on the GRE is their possessing the distinctive combination of mental characteristics that they do.
In that case, you are not really helping someone who is not constitutionally cut out for Philosophy by encouraging them to major in it anyway, in the hopes of improving their cognitive skills.
More to the point, it’s just bad thinking to draw that inference from these data.
Don’t get me wrong, I’m all for increasing the number of Philosophy majors.
But I don’t think Philosophers should continue to present a bad argument for this obviously laudable conclusion.
That just defeats the purpose of what we do.Report

Ben
Ben
Reply to  Wayne Fenske
1 year ago

I agree. We don’t know the causal explanation for why philosophy majors do well on standardized tests, but if I were to make a bet I’d put my money on its being in very large part a selection effect (philosophy selects for a certain kind of person who does well on the GRE). All my experience with philosophy majors suggests to me that this is the case.Report

grad student
grad student
Reply to  Ben
1 year ago

As I mentioned above, it would not just be a selection effect because philosophy selects for a certain kind of person, but the even stronger effect that philosophy *graduate school* selects for a certain kind of person.Report

Tomas Bogardus
Reply to  Wayne Fenske
1 year ago

I was pretty scrupulous in avoiding drawing any inferences, because it’s a very tricky topic! But I’ll go out on a limb here and say I think these graphs are evidence that Philosophy improves one’s ability to think critically. Because what if philosophers performed the very *worst* of any major on the GRE? I’d think that would count as evidence against the claim that Philosophy improves critical thinking. But then shouldn’t it follow that, if philosophers perform the very best of any major in the GRE, then this is evidence that Philosophy improves critical thinking?

Let’s call the hypothesis that Philosophy improves one’s ability to think critically, “H.” And the hypothesis that it doesn’t, “~H.” And the evidence that philosophers perform the very worst on the GRE, “E.” And the evidence that they don’t perform the very worst, “~E.”

John Hawthorne, Matthew Benton, and Yoaav Isaacs (2016) tell me this is a theorem of the probability calculus:

P(H|~E) > P(H) iff P(~H|E) > P(~H)

If you look at that right bijunct, that seems true in our case: the probability of ~H (the hypothesis that Philosophy doesn’t improve one’s ability to think critically) on E (the evidence that philosophers perform the very worst on the GRE) is greater than the intrinsic probability of ~H. If philosophers really performed the worst, that would be evidence that Philosophy doesn’t help critical thinking. (I think this might be the weakest link in this argument… But it seems to have a lot of truthiness, does it not?)

So, then, I suppose the right bijunct follows: the probability of H (that Philosophy *does* improve one’s ability to think critically) on evidence ~E (that philosophers do *not* perform the worst on the GRE) is greater than the intrinsic probability of H.

So that means that ~E raises the probability of H. Which means that the evidence I’ve shared (that philosophers don’t perform the worst on the GRE) is evidence that Philosophy improves one’s ability to think critically. A fortiori when the evidence is even stronger, namely that philosophers perform the very *best* on the GRE! 🙂

Feedback welcome! There’s quite possibly a mistake somewhere up there.Report

Tomas Bogardus
Reply to  Tomas Bogardus
1 year ago

Sorry, I meant to say that my last comment was directed at Wayne Fenske. 🙂Report

Tomas Bogardus
Reply to  Tomas Bogardus
1 year ago

Also, whoops, I wrote “I suppose the right bijunct follows” above, but I meant to say “I suppose the LEFT bijunct follows.” My bad. :-/Report

grad student
grad student
Reply to  Tomas Bogardus
1 year ago

Suppose there is a selection effect, and call that proposition S. Then it’s true that P(~H|E&S) > P(~H), since if the top philosophy students were the ones taking the GRE, and they did poorly, then surely that’s evidence that philosophy does not improve critical thinking. But from the theorem you cite does it follow that P(H|~E&S) > P(H)? I don’t see why immediately, though maybe I’m wrong. What immediately follows is P(H|~(E&S)) > P(H). But intuitively no, since both (1) they didn’t test the worst and (2) only the best were tested don’t jointly seem to support that philosophy improves your critical thinking.Report

Mathieu Doucet
Mathieu Doucet
Reply to  Wayne Fenske
1 year ago

Just because these data don’t show that a philosophy degree *causes* students to do well does not mean it’s not useful or interesting data, or not worth compiling. If the criterion for gathering and reporting on data in the social sciences was ‘demonstrates a clear causal relationship’, we’d have much less social science research.
I used to be our department’s undergraduate advisor. I had many conversations with students who worried (in part no doubt because their parents worried) that while philosophy was interesting, it wasn’t *useful*, was a waste of time, and would never lead to a good job. And it was always nice to be able to point to data like this to say, effectively, ‘students who study philosophy tend to do pretty well. Don’t worry that you’d be consigning yourself to a life of chronic underemployment.’
Maybe it’s all selection effect. (It’s definitely *partly* selection effect!) Maybe no undergraduate program actually teaches anyone any useful skills. But even so: if you’re the kind of student who is interested in studying philosophy but who worries about employability, you probably have the skills that the selection effect is selecting for. So study philosophy if you want to.
I’d be very upfront with students that the data didn’t show any causal relationship and that it might all be selection effect, and they both understood and (I think) appreciated this point. They also saw the data as useful. Report

Stefan Schubert
Stefan Schubert
1 year ago

Why use average standard deviations above the mean instead of the average official total score (at which philosophers didn’t end up top)? Of course any score can be objected against, but the average official score is of obvious intrinsic interest. It’s less clear why this score is interesting (and why it’s more interesting than the official score).Report

Tomas Bogardus
Reply to  Stefan Schubert
1 year ago

Hi Stefan. 🙂 My thought was this: the different subsections have different standard deviations. That means some sections are “harder” than others, i.e. getting a high score on some sections puts you in a higher percentile than getting that very same score on other sections. https://www.mbacrystalball.com/gre/gre-score-percentiles

But then it could happen that two students (or groups of students) get the same total score across these different sections, even though one student (or group of students) managed to do that by getting higher scores on the “harder” sections. I’d think that, in a case like that, we should be more impressed by this student (or group).

My understanding is that z-scores (number of SDs from the mean) can let us measure this. If you look at the data table provided above, you’ll see that Physics&Astronomy and Philosophy have *almost* the same total score when you add up the three sections. But Philosophy managed to do this by getting more impressive scores on the “harder” sections. Significantly more impressive. So hooray for philosophers, and that’s worth noting. 🙂

But, I’m a total amateur at stats, so I could well be wrong about this! Report

Stefan Schubert
Stefan Schubert
Reply to  Tomas Bogardus
1 year ago

This effectively means that one puts more weight on some sub-tests, and less on some sub-tests, than the test developers did. Presumably they had some reason for their weights. As stated, of course their relative weights could be criticised, as could any other relative weights, such as yours. But theirs is intrinsically interesting, because it is the official score which is used by, e.g. universities.Report

Ben
Ben
Reply to  Stefan Schubert
1 year ago

That’s not correct. The total GRE score has no particular meaning or significance from the point of view of the test constructors. It’s the subtests that are normed and scaled. Report

Ben
Ben
Reply to  Ben
1 year ago

…and reported. Thus, if you want a composite score, you need to create it yourself. Bogardus’s way of doing it makes at least as much sense as simply adding up the scores on the subtests.Report

Ben
Ben
Reply to  Tomas Bogardus
1 year ago

I think this approach makes sense. However, a major reason why the quantitative portion of the GRE has a lower ceiling than the verbal portion is not because it’s “objectively” easier, but simply because a huge number of non-native English speakers take the test.Report

clever
clever
1 year ago

Does anyone have recent data on annual salaries and salary growth by major? Particularly in an easily readable chart for including on a philosophy dept. poster?Report